Sensor For apnea classification and detection
Introduction : We are proposing an evolution of a sensor for apnea detection and classification. This sensor perceives pressure variations through the skin that indicate the presence or absence of respiratory effort during apneas. Moreover, the sensor detect apnea from respiratory sounds from tracheal sounds in the frequency band 200Hz – 2000Hz.
Method, Material : The sensor is a significant evolution of a sensor studied in [Meslier, 2002] (PNEAVOX Technology).
The former sensor was made with electrets used in CID102 (CIDELEC, France), in the frequency band 0,1Hz – 10Hz and gave a good specificity 93,6% and sensitivity 99,4% to classify apnea as obstructive, mixed or central. This method induced a delay in the signal that reduced the potential for apnea classification especially for children. To reduce the delay, we used a specific pressure sensor for respiratory effort detection and signal is filtered in the frequency band 0.02Hz-20Hz.
We present a mechanical model of two dimensions and three degrees of freedom with thoracic, abdomen and neck to represent dynamic of sensor signal. Pneavox measurements were performed in polysomnograph CID102L8D and CIDLX (Cidelec, France) to evaluate the initial results in patients.
Results : First tests shows good potential for apnea classification and detection. Introduction of the specific pressure sensor removes signal delay and so improves the technology dynamic for short event detection (children diagnosing). Moreover, examples show that the method indicates respiratory efforts when belts are flat during obstructive apneas [Boudewyns, 1997], and so potentially resolves the lack of belt sensitivity. Mechanical modelling explains interactions between respiratory efforts and actions around the neck (muscles …).
Conclusions : PneaVox technology is an insteresting solution for sleep diagnostic, for adults and also for children as it perceives respiratory efforts and breathing sounds with a good dynamic and without sensors placed on patient’s face.
Acknowledgement : We would like to acknowledge the CHU d’Angers (France), Charité Berlin (Germany), Hopital Necker (Paris France) for their help in testing material and improving sleep diagnostic tools for patients.